ArTailor

• 

  Website Interface Preview

Inspiration

We were inspired by controversies over the use of AI for art and GenAI threatening human artists. We wanted to be able to show that AI can and should be used as a tool to help artists, instead of entirely replacing them.

What it does

Our intention is to use ControlNet software to help artists compare their drawings' anatomy to their references. Besides ControlNet image processing, we also plan on this website consolidating other tools to help artists, such as color palette generators and color matchers.

How we built it

We used ControlNet's auxiliary models OpenPose for pose, hand, and face estimation/"pose tailoring," LineArt for generating sketchy-line-art, and MLSD line segment detection for estimating lines of perspective. We also have an additional Python program for comparing images with a back-and-forth slider.

Challenges we ran into

It was difficult at first to figure out how to allow users to upload files, but we were able to overcome that and provide options for URLs/direct uploads.

Accomplishments that we're proud of

We're proud of the overall design of the "website" and its intended user interface. We unfortunately weren't able to get the actual website working as we spent most of our time figuring out the pose estimation, but we were able to plan out how we would like the user interface to appear.

What we learned

We learned a lot about writing code in general. All four of us are relatively new to Hackathons and creating independent projects, so this was a nice learning experience overall.

What's next for ArTailor

We are thinking of fully developing this project in the future with more time, as we are genuinely passionate about using AI as a tool for artists rather than something to replace them entirely.

Built With

• controlnet
• genai
• google-colab
• lineart
• mlsd
• openpose
• python